Immediate versus delayed induction of labour in hypertensive disorders of pregnancy: a systematic review and meta-analysis

Background Mothers with hypertensive disorder of pregnancy can be managed with either immediate or delayed induction of labour with expectant monitoring of both mother and baby. There are risks and benefits associated with both the type of interventions. Hence, this review was conducted to compare outcomes of immediate and delayed induction of labour among women with hypertensive disorder of pregnancy based on disease severity and gestational age. Methods We conducted systematic searches in various databases including Medline, Cochrane Controlled Register of Trials (CENTRAL), Scopus, and Embase from inception until October 2019.Cochrane risk of bias tool was used to assess the quality of published trials. A meta-analysis was performed with random-effects model and reported pooled Risk ratios (RR) with 95% confidence intervals (CIs). Results Fourteen randomized controlled trials with 4244 participants were included. Majority of the studies had low or unclear bias risks. Amongst late onset mild pre-eclampsia patients, the risk of renal failure was significantly lower with immediate induction of labour (pooled RR: 0.36; 95%CI: 0.14 to 0.92). In severe pre-eclampsia patients, immediate induction of labour significantly reduced the risk of having small-for-gestational age babies compared to delayed induction of labour (pooled RR: 0.49; 95%CI: 0.29–0.84).Delayed induction was found to significantly reduce the risk of neonatal respiratory distress syndrome risk among late onset mild pre-eclampsia patients (pooled RR: 2.15; 95%CI: 1.14 to 4.06) None of the other outcomes demonstrated statistically significant difference between the two interventions. Conclusion Delayed induction of labour with expectant monitoring may not be inferior to immediate induction of labour in terms of neonatal and maternal outcomes. Expectant approach of management for late onset mild pre-eclampsia patients may be associated with decreased risk of neonatal respiratory distress syndrome, while immediate induction of labour among severe pre-eclampsia patients is associated with reduced risk of small-for-gestational age babies and among mild pre-eclampsia patients, it is associated with reduced risk of severe renal impairment. Supplementary Information The online version contains supplementary material available at 10.1186/s12884-020-03407-8.


Background
Hypertensive disorders during pregnancy are seen in approximately 3-10% of all pregnant women and can lead to serious maternal as well as neonatal morbidity and mortality [1][2][3].Globally, around 80-120 women die every day because of hypertensive complication during pregnancy [4].Mothers affected by the disorder are at higher risk of developing eclampsia with seizures, renal failure, liver failure, difficulty in breathing and HELLP syndrome (Haemolysis, elevated liver enzymes and low platelet count) [5][6][7][8].
Immediate delivery of the placenta remains the only definitive management for this condition. However, delivery of preterm baby is associated with increased perinatal mortality and other short and long-term perinatal complications like respiratory distress syndrome (RDS), neonatal seizures, intracerebral haemorrhage etc. [9][10][11][12][13]. Another area of concern is that women with immediate induction of labour have higher caesarean section rates [14].
Alternatively, the disorder can be managed by expectant monitoring of mother and baby and delaying the labour. Expectant management consists of frequent monitoring of blood pressure, maternal symptoms (like headache, abdominal pain, blurring of vision, decrease in foetal movements, vaginal bleeding), liver & renal function test and complete blood count depending on the severity of the disease. Indications for delivery of these women include poor blood pressure control despite treatment, developing eclamptic maternal symptoms like headache, visual disturbances, epigastric pain, nausea and vomiting, unfavourable blood tests result and decrease in foetal movements. However, this line of management is associated with increased maternal morbidity and mortality wherein the disease may progress to preeclampsia or eclampsia, HELLP syndrome or abruptio placenta [15,16].
In addition, maternal and neonatal outcomes with wither modes of intervention may vary depending on the severity of hypertension and gestational age [17][18][19]. American College of Obstericians and Gynecologists (ACOG) taskforce bulletin has indicated that a woman with pre-eclampsia should be delivered at 37 weeks of gestation [19]. Though, preterm delivery is considered as an option for severe pre-eclampsia, patients should be carefully evaluated for the adverse neonatal outcomes associated with the immediate induction of labour. Also, controversy exists on the benefits of having an elective delivery over delayed induction before34 weeks of gestation. ACOG taskforce bulletin has stated that "continued pregnancy can be undertaken at facilities only if there are adequate intensive care facilities for both mother and neonate at less than 34 weeks of gestation". It does not provide further instructions on to whether an elective delivery needs to be performed if complications like oliguria or anuria, swelling of feet, pulmonary edema and cerebral or visual disturbances occurs [19]. Despite these recommendations, several clinicians consider that delayed induction of labour as pre-eclampsia intervention after the 34 weeks of gestation will promote a better outcome for both mothers and neonates. However, evidences supporting the management criteria are very limited.
As both the immediate and delayed induction have their own advantages and disadvantages, there is a need for high-level evidence to assess which intervention results in better outcomes for hypertensive disorder during pregnancy. Hence, the purpose of this meta-analysis is to compare maternal and neonatal outcomes following immediate induction of labour vs delayed induction for hypertensive disorder of pregnancy based on severity and gestational age.

Inclusion criteria
We included only randomized controlled trials (RCTs) for the current review. The inclusion criteria were as follows: studies on women with hypertensive disorder of pregnancy irrespective of disease severity and gestational age; studies comparing immediate induction of labour and delayed labour with expectant monitoring of mother and baby; studies outcomes including any of the following.

Foetal/neonatal outcomes
Stillbirth rate (death of foetus at or after 28 weeks of gestation), perinatal mortality (death of foetus after 28 weeks of gestation till 7 days after delivery), neonatal mortality (death within 28 days of birth), neonatal morbidity (respiratory distress syndrome, neonatal seizures, small for gestational age, neonatal intensive care unit admission, necrotizing enterocolitis, intraventricular haemorrhage).
Non-randomised studies, single arm studies, abstracts, review articles were excluded.

Search strategy
Relevant articles for this study were identified by searching Medline, Embase, Cochrane Central Register of Controlled Trials (CENTRAL), Scopus, ScienceDirect and Google scholar databases and search engines for studies conducted from the inception (January 1964) to October 2019. Only papers published in English were included. Trial registries such as WHO International Clinical Trials Registry Platform (ICTRP) and ClinicalTrials. gov were also searched The following medical subject headings (MeSH) were used in combination with the free text terms: "pregnancy induced hypertension" OR "hypertensive disorder of pregnancy" OR "gestational hypertension" OR "pre-eclampsia" OR "hypertension" AND ("immediate delivery" OR "immediate induction of labour" OR "induction of labour") AND ("expectant management" OR "delayed induction of labour") AND "pregnancy" OR "pregnant women", with the limits to 'human' subjects and study design 'randomized controlled trial'. .
All relevant studies were analyzed separately by two reviewers based on the inclusion criteria listed above. The analysis was done first at the title and abstract level and then at the full-text level. Any disagreement was resolved by discussion with a third reviewer. Additionally, reference list of full-text articles was searched for any missed-out studies. The Preferred Reporting Items for Systematic Review and Meta-Analysis (PRISMA) check list was used for reporting the current review [17].

Data collection
The data extracted from the included studies contained all the details necessary for quality of study assessment, including: title and authors, study design and setting, participants, sample size, inclusion and exclusion criteria, characteristics of intervention and comparison groups, duration of follow up, primary and secondary study outcomes. Corresponding authors were contacted by email for missing data or when clarification or additional information was required for the methodological assessment of the included studies.
Outcome data were independently extracted by primary and secondary authors. In case of studies reporting multiple arms in a single trial, only the relevant arms were included. Collected data were transferred into the statistical software RevMan (ver 5.3) by the first author, and double checked by the third author.

Risk of bias assessment
The risk of bias for included studies was independently assessed by two authors using Cochrane risk of bias tool for RCTs [18]. The risk of bias was assessed by: generation of random sequence, allocation concealment, blinding of the participants and outcome assessment, incomplete outcome data and selective outcome reporting.
Risk of bias was graded based on the scoring of abovementioned domains; Grading was done as low, high or unclear based on the adequacy of information and satisfaction of the criteria.

Statistical analysis
Meta-analysis was performed using RevMan 5.3 (Copenhagen: The Nordic Cochrane Centre, The Cochrane Collaboration, 2014). The pooled effect for the dichotomous outcomes was estimated by retrieving number of events and number of participants for each study group to calculate Relative Risk-RR. Meta-analysis was performed separately for studies with mild to moderate pre-eclampsia and severe pre-eclampsia as defined by the individual studies. Further subgroup analysis under each of these domains was carried out based on the gestational age of onset (i.e. early or late onset). Random-effects model with inverse variance was used. In case of missing data, the corresponding author of the included trial was contacted. Alternatively, imputation methods were used to fill in the missing data. Heterogeneity and inconsistencies among the included studies were assessed by chi square test and I 2 statistics respectively. Results were graphically represented by forest plot. Selected studies were assessed for the reporting bias was assessed by comparing list of outcomes in the full study protocol with that of the published trial, where available.

Characteristics of included studies
Characteristics of the included studies are described in Table 1. All the included studies were RCTs. Most the studies (6 out of 14) were conducted in Europe. In total, 2132 participants in the immediate induction arm were compared with 2112 participants in the delayed induction arm. Total sample size of the studies varied from 30 to 946 while sample size in immediate induction arm varied from 15 to 471 and in delayed induction arm varied from 15 to 475. The definitions used for the diagnosis of gestational hypertension, pre-eclampsia and eclampsia are more or less similar across the studies included. However, there is a difference in the inclusion criteria based on the gestational age of onset and severity of the condition. Seven studies [15,[22][23][24][27][28][29] were conducted among late onset pre-eclampsia and rest of the studies [16,25,26,[30][31][32][33] among early onset preeclampsia. Eight studies were conducted on mild to moderate pre-eclampsia patients [15,16,22,23,[26][27][28][29].Six studies were conducted on severe pre-eclampsia patients [24,25,[30][31][32][33].

Quality of the included studies
Authors assessment of risk of bias for included studies is presented in Table 2. Most of the studies had low risk of bias with respect to the randomization process (random sequence generation and allocation concealment). All the included studies had either high or unclear risk of bias with respect to blinding of participants and outcome assessment. All the studies had low or unclear risk of bias with respect to incomplete outcome data and selective reporting of outcome except Majeed et al. [29] and Mesbah EMM [30].

Outcomes
All the outcome estimates based on the gestational age of onset and severity of condition is provided in the Table 3.

Maternal outcomes Maternal mortality
Seven studies reported data on maternal mortality. Five of those studies had zero incidence of maternal mortality in both study groups Stratification based on severity and gestational age of onset had one study each under mild [15] and severe pre-eclampsia [24] and both were conducted among late onset pre-eclampsia patients.
Three studies reported data on the risk of postpartum haemorrhage. All the three studies were conducted among late onset mild pre-eclampsia patients. Metaanalysis indicated no statistically significant difference in risk of postpartum haemorrhage between the two groups [RR:0.82 (95%CI: 0.56 to 1.19); I 2 = 0%] (Additional file 2: Appendix 1.10). All the 14 studies have reported on the incidence of caesarean section among mothers in both the arms. The pooled RR for late onset mild pre-    Ten studies have reported on the neonatal intensive care unit (NICU) admission in both the arms. Six studies were conducted among mild pre-eclampsia patients (one early onset and five late onset preeclampsia patients). The pooled RR for late onset patients was 1.28 (95%CI: 0.87 to 1.87; I 2 = 10%) (Additional file 2: Appendix 1.30) and RR for early onset patients was 4.00 (95%CI: 1.23-13.05, I 2 = NA) (Additional file 2: Appendix 1.31). Four studies were conducted among early onset severe pre-eclampsia patients. The pooled RR was 1.22 (95%CI: 0.95-1.56; I 2 = 75%) (Additional file 2: Appendix 1.32). Two studies have reported on intraventricular haemorrhage in both the arms (one study among early onset severe pre-eclampsia patients and other in late onset mild pre-eclampsia patients). The RR reported in early onset severe patients were 4.03 (95%CI: 0.46-35.59; I 2 = Not applicable) and early onset mild patients were 1.82 (1.06-3.14; I 2 = Not applicable). Four studies have reported on necrotizing enterocolitis, out of which three were conducted among early onset severe patients and only one among early onset mild patients. The pooled RR for early onset severe patients were 2.23 (95%CI: 0.42-11.87; I 2 = 29%) (Additional file 2: Appendix 1.33).

Discussion
Mothers with hypertensive disorder of pregnancy can be managed with either immediate induction of labour or delayed induction with expectant monitoring of both mother and baby. There are risks and benefits associated with both the type of interventions. Hence, it is important to analyse literature so as to which intervention is associated with better maternal and foetal outcomes.
In all, we identified 14 studies with 4244 participants for our analysis. Majority of the studies were conducted in countries of European region and American region. Most of the included studies had low risk of bias with respect to all the domains except blinding process domains. We did not find any substantial heterogeneity for most of the outcomes in the studies except studies reporting caesarean section rate in both the arms.
Only three outcomes (one maternal and two neonatal outcomes) showed a statistically significant difference between the two interventions. Immediate induction of labour was beneficial to mothers as it had reduced risk of severe renal impairment among late onset mild pre-eclampsia patients. It was also beneficial in neonatal outcome in terms of SGA babies among severe pre-eclampsia patients. Our analysis also demonstrated that delayed induction significantly reduces the risk of neonatal respiratory distress syndrome among late onset mild preeclampsia patients. Our results are similar to the previous Cochrane review conducted by Churchill D et al. in 2018 [34]. In their study too, immediate induction of labour was associated with reduced risk of SGA babies and the risk of neonatal respiratory distress syndrome was found to be lower with delayed induction and expectant management. In our results, no statistically significant difference was noted in any of the other outcomes as the confidence interval crossed the null value in all our remaining analyses. This indicates that evidence on the superiority of one intervention over the other for managing women with hypertensive disorder of pregnancy is limited.
The ACOG guidelines classify mild pre-eclampsia as systolic blood pressure of 140-149 mmHg and/or diastolic pressure of 90-99 mmHg, in women with a previously normal blood pressure with proteinuria of ≥300 mg/24-h urine collection. On the other hand, severe pre-eclampsia is classified as systolic blood pressure of > 160 mmHg and/or diastolic pressure of > 110 mmHg with severe proteinuria (2-5 g/24-h urine collection) [19]. It is important to note that majority of the include studies in our analysis were conducted on early onset severe pre-eclampsia and late onset mild pre-eclampsia. The number of studies pooled for a meta-analysis on maternal and neonatal outcomes for early onset mild pre-eclampsia were limited to just two. While our analysis did not find any difference between the two interventions for incidence of caesarean section, perinatal mortality and neonatal mortality with pooled analysis of the two studies, lack of data significantly limits the ability of our review to derive strong conclusions for this sub-group. In view of this, clinicians should carefully weigh the risk vs benefits of each intervention on a case-to-case basis to achieve optimal maternal and fetal outcomes for patients with early mild pre-eclampsia.
Three other similar reviews have been conducted on this topic, a meta-analysis of RCTs by Cluver C et al.  [35][36][37]. However, the number of studies included in these previous reviews have been limited to only 4-5 trials. In comparison, the current review has analysed 14 studies and comprehensively reviewed the maternal and neonatal morbidity & mortality outcomes with immediate induction and delayed induction of labour. We believe, the updated evidence shall help healthcare personnel in appropriately choosing the timing of induction of labour for the patients with hypertensive disorder of pregnancy.
The major strengths of our study include the comprehensive search of literature and the broad search strategy resulting in inclusion of all studies published on the topic to date. We only included RCTs into our review which enables us to infer causal associations between the intervention and outcomes. Several maternal and neonatal outcomes were analysed in our study thereby providing a comprehensive comparison of the two intervention modalities. The level of inter-study heterogeneity among majority of the maternal or neonatal outcomes was low. Outcomes were stratified based on the disease severity and gestational age of onset to take into account the influence of such confounding factors on the overall outcome.
We are also aware of the limitations of our review. We could not assess the possibility of publication bias for majority outcomes due to the limited number of included studies. Finally, most of the studies included in our review were conducted in European region, which may limit the generalizability of our findings to other geographical areas.

Conclusion
To summarize, delayed induction of labour with expectant monitoring may not be inferior to immediate induction of labour in terms of maternal and neonatal outcomes. Expectant approach of management for late onset mild pre-eclampsia patients may be associated with decreased risk of neonatal respiratory distress syndrome. Immediate induction of labour among severe pre-eclampsia patients is associated with reduced risk of SGA babies and among mild pre-eclampsia patients, it may be associated with reduced risk of severe renal impairment. Evidence on early onset mild pre-eclampsia was limited to draw strong conclusions.